Useful R functions

← Flow control (if/else, for) Self-study assignment →

Sandbox vectors

Let’s define some vectors which can be used for demonstrations:

manyNumbers <- sample( 1:1000, 20 )
manyNumbers

 [1] 294 507 756 861 986 461   8 826 597 365 240 256 517 157 471 795 871 403 982 868

manyNumbersWithNA <- sample( c( NA, NA, NA, manyNumbers ) )
manyNumbersWithNA

 [1] 795 471 256 403 756 294 982 240 871 986 517 861  NA 157  NA   8  NA 461 507 826 365 868 597

duplicatedNumbers <- sample( 1:5, 10, replace = TRUE )
duplicatedNumbers

 [1] 5 5 4 3 4 5 2 5 3 5

letters

 [1] "a" "b" "c" "d" "e" "f" "g" "h" "i" "j" "k" "l" "m" "n" "o" "p" "q" "r" "s" "t" "u" "v" "w" "x" "y" "z"

LETTERS

 [1] "A" "B" "C" "D" "E" "F" "G" "H" "I" "J" "K" "L" "M" "N" "O" "P" "Q" "R" "S" "T" "U" "V" "W" "X" "Y" "Z"

mixedLetters <- c( sample( letters, 5 ), sample( LETTERS, 5 ) )
mixedLetters

 [1] "u" "s" "n" "k" "z" "B" "W" "Q" "U" "A"

Are all/any elements TRUE

• Input: logical vector
• Output: single logical value
• Task: try, understand what happens when you use manyNumbersWithNA instead of manyNumbers.

all( manyNumbers <= 1000 )

[1] TRUE

all( manyNumbers <= 500 )

[1] FALSE

any( manyNumbers > 1000 )

[1] FALSE

any( manyNumbers > 500 )

[1] TRUE

all( !is.na( manyNumbers ) )

[1] TRUE

any( is.na( manyNumbers ) )

[1] FALSE

Which elements are TRUE

Input: logical vector Output: vector of numbers (positions)

which( manyNumbers > 900 )

[1]  5 19

which( manyNumbersWithNA > 900 )

[1]  7 10

which( is.na( manyNumbersWithNA ) )

[1] 13 15 17

Filtering vector elements

• Input: any vector and filtering condition
• Output: elements of the input vector
• Note: several ways to get the same effect

manyNumbers[ manyNumbers > 900 ] # indexing by logical vector

[1] 986 982

manyNumbers[ which( manyNumbers > 900 ) ] # indexing by positions

[1] 986 982

somePositions <- which( manyNumbers > 900 )
manyNumbers[ somePositions ]

[1] 986 982

Are some elements among other elements

• Input: two vectors
• Output: a logical vector corresponding to the first input vector

"A" %in% LETTERS

[1] TRUE

c( "X", "Y", "Z" ) %in% LETTERS

[1] TRUE TRUE TRUE

all( c( "X", "Y", "Z" ) %in% LETTERS )

[1] TRUE

all( mixedLetters %in% LETTERS )

[1] FALSE

any( mixedLetters %in% LETTERS )

[1] TRUE

mixedLetters[ mixedLetters %in% LETTERS ]

[1] "B" "W" "Q" "U" "A"

mixedLetters[ !( mixedLetters %in% LETTERS ) ]

[1] "u" "s" "n" "k" "z"

manyNumbers %in% 300:600

 [1] FALSE  TRUE FALSE FALSE FALSE  TRUE FALSE FALSE  TRUE  TRUE FALSE FALSE  TRUE FALSE  TRUE FALSE FALSE
[18]  TRUE FALSE FALSE

which( manyNumbers %in% 300:600 )

[1]  2  6  9 10 13 15 18

sum( manyNumbers %in% 300:600 )

[1] 7

Pick one of two (three) depending on condition

• Input: a logical vector and two vectors additional vectors (for TRUE, for FALSE)
• Output: elements of the additional vectors
• Note: it can take care of NAs

if_else( manyNumbersWithNA >= 500, "large", "small" )

 [1] "large" "small" "small" "small" "large" "small" "large" "small" "large" "large" "large" "large" NA     
[14] "small" NA      "small" NA      "small" "large" "large" "small" "large" "large"

if_else( manyNumbersWithNA >= 500, "large", "small", "UNKNOWN" )

 [1] "large"   "small"   "small"   "small"   "large"   "small"   "large"   "small"   "large"   "large"  
[11] "large"   "large"   "UNKNOWN" "small"   "UNKNOWN" "small"   "UNKNOWN" "small"   "large"   "large"  
[21] "small"   "large"   "large"  

# here integer 0L is needed instead of real 0.0 
# manyNumbersWithNA contains integer numbers and the method complains
if_else( manyNumbersWithNA >= 500, manyNumbersWithNA, 0L ) 

 [1] 795   0   0   0 756   0 982   0 871 986 517 861  NA   0  NA   0  NA   0 507 826   0 868 597

Duplicates and unique elements

• Input: a vector

unique( duplicatedNumbers )

[1] 5 4 3 2

unique( c( NA, duplicatedNumbers, NA ) )

[1] NA  5  4  3  2

duplicated( duplicatedNumbers )

 [1] FALSE  TRUE FALSE FALSE  TRUE  TRUE FALSE  TRUE  TRUE  TRUE

Positions of max/min elements

which.max( manyNumbersWithNA )

[1] 10

manyNumbersWithNA[ which.max( manyNumbersWithNA ) ]

[1] 986

which.min( manyNumbersWithNA )

[1] 16

manyNumbersWithNA[ which.min( manyNumbersWithNA ) ]

[1] 8

range( manyNumbersWithNA, na.rm = TRUE )

[1]   8 986

Sorting/ordering of vectors

manyNumbersWithNA

 [1] 795 471 256 403 756 294 982 240 871 986 517 861  NA 157  NA   8  NA 461 507 826 365 868 597

sort( manyNumbersWithNA )

 [1]   8 157 240 256 294 365 403 461 471 507 517 597 756 795 826 861 868 871 982 986

sort( manyNumbersWithNA, na.last = TRUE )

 [1]   8 157 240 256 294 365 403 461 471 507 517 597 756 795 826 861 868 871 982 986  NA  NA  NA

sort( manyNumbersWithNA, na.last = TRUE, decreasing = TRUE )

 [1] 986 982 871 868 861 826 795 756 597 517 507 471 461 403 365 294 256 240 157   8  NA  NA  NA

manyNumbersWithNA[1:5]

[1] 795 471 256 403 756

order( manyNumbersWithNA[1:5] )

[1] 3 4 2 5 1

rank( manyNumbersWithNA[1:5] )

[1] 5 3 1 2 4

sort( mixedLetters )

 [1] "A" "B" "k" "n" "Q" "s" "u" "U" "W" "z"

Ranking of vectors

manyDuplicates <- sample( 10:15, 10, replace = TRUE )
rank( manyDuplicates )

 [1] 4.5 6.5 8.0 9.5 2.0 6.5 4.5 2.0 9.5 2.0

rank( manyDuplicates, ties.method = "min" )

 [1] 4 6 8 9 1 6 4 1 9 1

rank( manyDuplicates, ties.method = "random" )

 [1]  4  7  8  9  2  6  5  1 10  3

Rounding numbers

v <- c( -1, -0.5, 0, 0.5, 1, rnorm( 10 ) )
v

 [1] -1.00000000 -0.50000000  0.00000000  0.50000000  1.00000000 -0.41755004  0.42357074 -0.05794781
 [9] -0.31518870  0.79171579  1.32086709  0.03595770 -0.34807456 -0.85789025  0.23201075

round( v, 0 )

 [1] -1  0  0  0  1  0  0  0  0  1  1  0  0 -1  0

round( v, 1 )

 [1] -1.0 -0.5  0.0  0.5  1.0 -0.4  0.4 -0.1 -0.3  0.8  1.3  0.0 -0.3 -0.9  0.2

round( v, 2 )

 [1] -1.00 -0.50  0.00  0.50  1.00 -0.42  0.42 -0.06 -0.32  0.79  1.32  0.04 -0.35 -0.86  0.23

floor( v )

 [1] -1 -1  0  0  1 -1  0 -1 -1  0  1  0 -1 -1  0

ceiling( v )

 [1] -1  0  0  1  1  0  1  0  0  1  2  1  0  0  1

Naming vector elements

heights <- c( Amy = 166, Eve = 170, Bob = 177 )
heights

Amy Eve Bob 
166 170 177 

names( heights )

[1] "Amy" "Eve" "Bob"

names( heights ) <- c( "AMY", "EVE", "BOB" )
heights

AMY EVE BOB 
166 170 177 

heights[[ "EVE" ]]

[1] 170

Generating grids

expand_grid( x = c( 1:3, NA ), y = c( "a", "b" ) )

# A tibble: 8 x 2
      x y    
  <int> <chr>
1     1 a    
2     1 b    
3     2 a    
4     2 b    
5     3 a    
6     3 b    
7    NA a    
8    NA b    

Generating combinations

combn( c( "a", "b", "c", "d", "e" ), m = 2, simplify = TRUE )

     [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10]
[1,] "a"  "a"  "a"  "a"  "b"  "b"  "b"  "c"  "c"  "d"  
[2,] "b"  "c"  "d"  "e"  "c"  "d"  "e"  "d"  "e"  "e"  

combn( c( "a", "b", "c", "d", "e" ), m = 3, simplify = TRUE )

     [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10]
[1,] "a"  "a"  "a"  "a"  "a"  "a"  "b"  "b"  "b"  "c"  
[2,] "b"  "b"  "b"  "c"  "c"  "d"  "c"  "c"  "d"  "d"  
[3,] "c"  "d"  "e"  "d"  "e"  "e"  "d"  "e"  "e"  "e"  

← Flow control (if/else, for) Self-study assignment →

---

Copyright © 2021 Biomedical Data Sciences (BDS) | LUMC